The colonic microbiome has been implicated in colorectal cancer (CRC) pathogenesis but the exact mechanisms underlying these observations remain incompletely understood. We have observed striking associations between specific host microbes and aberrant DNA methylation in CRC. For example, Fusobacterium species are substantially enriched in cancers affected by the CpG Island Methylator Phenotype (CIMP). Preliminary data based on sequencing and qPCR validation also show enrichment of bacteria that have been linked to disease in humans or mouse models (E.coli sp., Klebsiella sp. etc.). High levels of these pathogenic bacteria are associated with recurrences in CIMP+ colon cancers. This unexpected link between colonic microbiota and epigenetic control was also seen in an analysis of DNA methylation genome wide in the colonic mucosa of germ free (GF) mice compared to conventionalized mice, where we found that reintroduction of bacteria led to hypermethylation of normally unmethylated CpG island sites (the main anomaly seen in CIMP+ cases). Thus, our preliminary data support a new hypothesis, that the GI microbiota affects colonic neoplasia through inducing or modulating aberrant DNA methylation and epigenetic control. Mechanistically, we propose that multiple parallel mechanisms may be contributing to this link including DNA damage associated recruitment of silencing complexes, and metabolic disturbances whereby bacteria secrete metabolites and/or toxins that diffuse into colonic epithelial cells and affect DNA methylation directly (e.g. 2-hydroxyglutarate [2-HG], which inhibits the TET DNA demethylase enzymes) or indirectly (e.g. butyrate, which is known to modulate epigenetics through inhibition of histone deacetylases). To test these hypotheses, we propose three specific aims: (1) Define the microbiome across the spectrum of CIMP+ tumors. We will use 16S RNA genomic sequencing in an extensive tumor (cancer, precursors and adjacent normal) set simultaneously characterized for CIMP, mutations and gene expression. (2) Impact of CIMP+ associated bacteria on tumorigenesis and DNA methylation in mice. GF Il10-/-;Apcmin/+ mice will be colonized with bacterial candidates (e.g. E.coli, F.nucleatum, K.pneumonia) and tumor incidence, severity, survival as well as DNA methylation and gene expression in normal and tumor tissues will be evaluated. We will also test whether drugs targeting DNA methylation are effective in prevention of bacteria-associated tumorigenesis. (3) Study metabolites by which bacteria influence DNA methylation profiles. We will use metabolomics on cultures of bacteria associated with CIMP and on lysates from CIMP+ and CIMP- cancers to identify metabolites that potentially modulate DNA methylation. These (e.g. 2HG, butyrate) will be tested for effects on DNA methylation (in cell culture and in GF mice) and tumorigenesis (in mice). The proposed research tests a new mechanism for microbiome-associated tumorigenesis and has important implications for detection, prevention and treatment of CRCs.